一种改进的SIR粒子滤波方法

摘要

sampling importance resampling ( SIR ) particle filtering may fail when the resampling process becomes invalid. The failure reason of SIR algorithm is firstly analyzed, and then an improved method is proposed to solve it. The improved method adopts partide swarm optimization ( PSO) algorithm to guide particles to move into a higher likelihood location to ensure that the resample process could run well. Simulation results show that the proposed method can work well to solve the filtering failure problem in the case aforementioned.%SIR粒子滤波算法在重采样无法进行时可能失效,详细分析了算法失效的原因,并针对此问题提出了基于PSO的改进方法,该方法利用PSO的智能寻优机制引导重要性抽样的粒子移向高似然区,从而确保重采样过程的顺利进行.仿真试验表明,提出的改进方法可以有效解决SIR算法因重采样无法进行而导致的失效问题.